What is ideal Function Generator for audio?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Bluzzi

Well-known member
Joined
Jun 6, 2004
Messages
1,229
Location
Montreal
I'm to the point where I need to get a Function Generator and looking at the hundred of models available its hard to make a decision without feeling like I may miss out on a feature that is not apparently required now.

Looks like the 2Mhz models are decently priced. I'd prefer one with a Frequency display.

How important is the distortion and stability of the waves? Some don't have level adjustment and I feel that is going to be necessary.

What about sweep frequency feature?. anything you can think off that comes in handy that is not standard on most models?

Curious to see the opinions of those who have had more experience with Function Generators than me (none).

jim
 
I designed a sinewave oscillator for general audio troubleshooting and measurement. While a function generator grade sinewave is adequate for frequency response measurements, for listening tests, I thought the .2% or so THD of mine was borderline high, when listening for speaker rubs, or general sound quality.

YMMV

JR
 
I need mine for accurate measurements for sure so yes I guess the lower distortion will be preferable.

jim
 
A couple tenths THD won't cause measurement errors for anything other than THD measurements where you want a very low distortion floor. For frequency response it's not a problem.

JR
 
True. I was thinking more along the lines of a visual distortion when displaying lets say a Square wave. If the source has a slight imperfection that the oscilloscope shows it then becomes hard to know which portion comes from source and which comes from circuit being tested.

But yes, if its something like .02 compared to .008 there isn't going to be that much to worry about.

What about drift over time is that a big deal? I see that spec on many units as though it is a big deal. Again it would seem to me that it should not be a huge worry in my case.

I'd love to test for precise distortion but I can't afford an Audio Precision unit.
Are there any labs that one can send units to get an independent test results?

jim
 
[quote author="Bluzzi"]True. I was thinking more along the lines of a visual distortion when displaying lets say a Square wave. If the source has a slight imperfection that the oscilloscope shows it then becomes hard to know which portion comes from source and which comes from circuit being tested.
jim[/quote]

The square waves are usually very clean on the stock mid-priced function generators. The waveforms are generated with a triangle/square wave engine to begin with, and the problem is taking the triangle and turning it into an approximate sinusoid. In the standard arrangement this is done not via filtering but rather a nonlinear network, usually using diodes and resistors.* Since the diodes have manufacturing variances and temp effects this is where, mostly, the variation arises among different units.

So the sine wave output is not very useful for any but gross distortion evaluations, like the onset of hard clipping. You might want to look for an old HP or other dedicated sine wave gen for finer measurements. These manage about 0.05% typically.


*I've found this fascinating as a counterexample to how most nonlinear networks add harmonic energy, rather than in this case removing it. Of course it is a very contrived situation, with the level dependence and all.
 
A used HP 3310 on ebay is an excellent bet. Very wide bandwidth, super stable and accurate, all the waveforms you need, reliable. Not expensive, highly recommended.
 
thanks all.

I'm compiling a list of models, prices and features to make a final decision.
your comments have helped me focus on the important aspects.

jim
 
If you are planning on doing distortion measurements, you will need a (non-function) sine wave generator. Function generators use sine approximation so they are limited in terms of producing a low distortion sine wave. There are some DIY low distortion generators floating around. Even good used sine wave generators like HP built have what is now considered high distortion at .01%. The Sound Technology 1700B distortion measuring set shows up on eBay from time to time. This unit can get down to about .004 or .005 on a good day.
 
I'm looking for some decent and especially not too expensive function generator (for testing built hardware) also possibly with sweep function (this is good to have, right?). Is this OK or is this some cheap crappy machine: http://www.amazon.de/Funktions-Generator-McCheck-FG-1641B-digi-1Hz-3MHz/dp/B000AKCAGC ? I might get it for reasonable price.
The other generator I might get is Philips PM5131. Which of these two is better for my purpose? I'm leaning toward FG1641 because it has display (good to know exact fq).
Thanks!
 
It depends on what your definition of 'testing' is. If you want to perform FR measurement and check for overshoot / square wave response, then this generator is likely to be fine. Personally I use an old function gen - I don't see the point in having an expensive one, as unless you spend a fortune it's unlikely to be much use in measuring distortion.

To reiterate what the previous posts say: an FG + scope will serve you well for 90%+ of testing scenarios, but if you want to get a little deeper in terms of THD measurement you'll need a dedicated low-distortion oscillator.

BTW - I can't see why an analogue FG that puts out '10.1Khz' is any less useful than a digital one that gives an accurate 10Khz. Unless you're measuring distortion there's little need for exact frequencies. 
 
> good to know exact fq

Why?

I like the H-P 200AB. Large, old, will need fixing. But an honest KNOB instead of these silly buttons.

The other tempting machine goes for $13 on eBay, a bare board (with buttons, yuk) about the size of a pack of Newport. Seems to do all the right stuff, and the price is great.
 
Hi,
Please look at Miniature Audio Oscillator : http://sound.westhost.com/project86.htm
"The Miniosc is not a toy oscillator. It is capable of serious work testing domestic or professional audio equipment of all types and will verify normal operation, allow levels to be set, channels to be matched and response curves measured."

Cheers,
P.
 
There are many, many uses for signal generators and that is probably why you're having a hard time choosing "only one". I use an Audio Precision 2322 for distortion and response tests, but I also have a Tektronix 191 for high frequency testing of capacitors and inductors, as well as an HP 8082A pulse generator for measurements more related to the time domain than the frequency domain. Why? Because the AP can't do all of that!

Of course, this is along with some nice 'scopes to see what you're doing. I'm a huge fan of the Tek 7000 series - the 7704 and 7904 are pretty awesome with the right plugins.

Oscillators suitable for distortion tests are gonna be not so cheap, but a simple audio band response sweep could be done with a soundcard and clever software. My suggestion is that "classic" American test gear, like that from HP and Tektronix, is probably not going to be a bad value, as long as it does something useful for you. One used to be able to find high quality T&M gear like this at amateur radio 'hamfests', but this stuff has largely dried up because of the magic of eBay. Still, depending on where you are, there might be good surplus gear, so it's worth digging a bit.

Just figure out what you need to do, find gear that does it, and don't be ashamed when you end up with several boxes that seem to do similar things - many times, they don't!
 
I do have a nice set of Hameg serie 8000 type 8027 + 8037
Perhaps you still can find them on the internet secondhands....
 
May be stupid, but why not use the output of a good soundcard ? For example, rme claims a 0.00032% thd for the output of their fireface 802...
 
Back
Top